Diagnosis of acute bacterial pneumoniain Nigerian children · (ii) bronchopneumonia, in which there was more widespread or ill-defined consolidation. The x-rays of one patient subsequently

Archives of Disease in Childhood, 1977, 52, 925-931

Diagnosis of acute bacterial pneumonia in NigerianchildrenValue of needle aspiration of lung and of countercurrent immunoelectrophoresis

M. SILVERMAN, D. STRATTON, A. DIALLO, AND L. J. EGLER

From the Departments ofPaediatrics and Microbiology, Ahmadu Bello University, Zaria, Nigeria

SUMMARY Eighty-eight Nigerian children with untreated, severe, acute pneumonia were investigatedby standard bacteriological techniques (blood culture and culture of pharyngeal secretions) and byneedle aspiration of the consolidated lung. Countercurrent immunoelectrophoresis (CIE) againstgrouped pneumococcal and Haemophilus influenzae type b antisera was carried out on serum samplesfrom 45 patients.The aetiology of pneumonia was shown by examination of the needle aspirate in 70/88 patients

(79 %), by CIE in 9/45 patients (20 %), and by blood culture in 4/36 patients (11 %). Overall, abacterial cause for pneumonia was shown in 73/88 patients (83 %). The results of pharyngeal culturewere misleading when compared with cultures of needle aspirates. The prediction of aetiology fromthe radiological appearance was also inaccurate, even for lobar pneumonia. Needle aspiration of thelung, with a low (5 %) and minor complication rate, merits wider application in the diagnosis ofacute pulmonary infections in children. Traditional bacteriological techniques (blood culture andpharyngeal culture) are of very limited value. The place of CIE in the inevstigation of childhoodpneumonia still needs thorough evaluation.

It is rarely possible to be certain of the cause of pneu-monia in a child. The traditional bacteriological in-vestigations of sputum examination and blood cul-ture each provide the correct answer in a minorityof patients and frequently confuse the issue with bothfalse-positive and false-negative results (Loda et al.,1968; Barrett-Connor, 1971; Kinnell, 1973; Glezenand Denny, 1973; Tugwell and Greenwood, 1974;Davidson et al., 1976). In developing nations, diseasepatterns may be affected by malnutrition, the absenceof widespread immunization programmes, and thelack of early medical attention (Morley, 1973). Pre-vious surveys in such areas have shown that, as inthe preantibiotic era in Western Europe and America(Rosenow, 1911; Lyon, 1922; Ellison, 1931; Sap-pington and Favorite, 1936; Butler et al., 1941),there is a high prevalence of bacterial infections ofthe lower respiratory tract (Abdel-Khalik et al.,1938; Klein, 1969; Mimica et al., 1971). The wide-spread and early use of antibiotics in more advancedsocieties may not only render more difficult thediagnosis of bacterial pneunmonia (Spencer and Philp,

Received 20 May 1977

1973) but may also have contributed to the appar-ently higher prevalence of mycoplasma and viralrespiratory tract infections in children (Nichol andCherry, 1967; Loda et al., 1968; Gardner, 1971).From time to time over the last 93 years, direct

needling of the lung (needle aspiration of the lung)has been advocated as an accurate means of deter-mining the aetiology of pneumonia (Leyden, 1883;Ellison, 1931; Alexander et al., 1941; Finland, 1969;Hyde et al., 1973), but there have been few thoroughassessments of the technique in children. Recently,countercurrent immunoelectrophoresis (CIE) has be-come available for the rapid identification of bac-terial antigens in patients with pneumonia (Lancet,1976). The value of the technique in childhood pneu-monia has not yet been established (Lampe et al.,1976; Michaels and Poziviak, 1976).Our study was set up to identify the bacteria

causing severe pneumonia in children living in thearea served by Ahmadu Bello University Hospital,using standard bacteriological techniques as well asneedle aspiration of the lung and CIE. The environ-ment consisted of a largely traditional Hausa town(Zaria) and its rural environs in the savanna belt of

925

on May 1, 2021 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.52.12.925 on 1 D

ecember 1977. D

ownloaded from

http://adc.bmj.com/

926 Silverman, Stratton, Diallo, and Egler

West Africa. By establishing the pattern of bacterialpneumonia in the area, we hoped to devise a standardform ofmanagement which could be applied withoutextensive investigation in the future.

Patients and methods

A total of 88 children with severe acute pneumoniawho were admitted to the paediatric emergency wardof Ahmadu Bello University Hospital, Zaria, werestudied. All were over 3 months of age and none hadbeen given antibiotics during their current illness;all had radiological evidence of pulmonary con-solidation. Ages ranged from 4 months to 8 yearsand a high proportion of the patients were under-weight or frankly marasmic (Table 1).

Table 1 Distribution ofpatients by age and weight

Age (yrs)

0-1 -2 -3 -4 -5 -6 -7 -8 -9 -10

No. of patients 23 20 18 4 4 7 6 4 1 1Underweight* 5 8 9 1 1 4 1 0 1 1Marasmict 4 4 6 2 2 1 2 1 0 0

*60.80% of Harvard standard weight-for-age. tLess than 60% ofHarvard standard weight-for-age (Nelson et aL, 1969).

All children had a clinical history, physical ex-amination, and a chest x-ray carried out. The radio-logical appearance was classified, independently ofbacteriological information, into two categories:(i) lobar pneumonia, in which consolidation con-formed to segmental (anatomical) boundaries, and(ii) bronchopneumonia, in which there was morewidespread or ill-defined consolidation. The x-raysof one patient subsequently disappeared withoutrecord. The presence of other features (pleuraleffusions, pneumatoceles) was noted.

In all patients needle aspiration of the lung wascarried out, the region of maximum consolidationhaving been determined by inspection of the chestx-ray. In performing needle aspiration the anteriorchest wall was avoided, to minimize anxiety and toavoid the possibility of entering the heart with theneedle. The skin was sterilized with iodine. A stan-dard 18 or 21 gauge bevelled needle attached to a10 ml syringe was inserted. As soon as the skin hadbeen pierced, a negative pressure was applied to thesyringe and the needle was plunged swiftly into thelung aiming towards the hilum, and removed withouta pause. Pressure was released just before withdrawalfrom the skin. The tiny drop of lung aspirate (up to0 5 ml) was diluted with 0 5 ml sterile nutrient brothbefore being used to prepare films for Gram stainingand for direct inoculation of standard chocolate and

blood agar plates. Colony identification was by stan-dard bacteriological techniques.Venous blood was taken for blood culture and

examination for bacterial antigen by CIE. CIE wascarried out with grouped pneumococcal antiserum(Omniserum, Statens Seruminstitut, Copenhagen)and grouped Haemophilus influenzae type b anti-serum (Hyland Laboratories) using a technique fullydescribed previously (Tugwell and Greenwood,1975). Finally, a pharyngeal swab was collectedduring stimulated coughing. The pharyngeal exudatewas cultured and bacterial colonies identified.

Patients were kept under close observation in theemergency ward, and further chest x-rays were takenif indicated. It was not possible to carry out follow-up chest x-rays as a routine procedure.

Results

Lung aspiration (Table 2). In every case lung fluid wasobtained. The amount varied from an immeasurablysmall drop to 0( 5 ml, and the consistency fromapparently pure blood to opaque yellow fluid. Anaetiological diagnosis was made in 70 of the 88patients (79 %): in 54 (61 %) by culture of organisms,and in a further 16 (18 %) by identification of organ-isms on the Gram-stained film of the aspirate whenculture was sterile. A single organism was isolatedin 52 patients, 2 organisms in 14 patients, and 3organisms in a further 4 patients.

Table 2 Organisms isolatedfrom lung aspirate

By culture By Gram stainof aspirate or culture(no. ofpatients) (no. ofpatients)

Pneumococci 31 45,3-haemolytic streptococci 3 3a-haemolytic streptococci 5 5Staphylococcus aureus 8 10Coliforms 7 14Haemophilus lnfluenzae 9 10Bacteroides spp. 2 2Flavobacterium 1 1Salmemhla-4yphi 1 1Candida albicans 1 1

No organism isolated 34 18

Note: Mixed infections occurred in 18 patients.

Complications of needle aspiration occurred in 4patients. 2 had transient minor haemoptyses, ex-pectorating small amounts of freshly blood-stainedsputum for about 3 minutes after the procedure. Onechild had a small pneumothorax which did not re-quire active treatment, and one had subcutaneousemphysema confined to the aspiration site.

There were 4 deaths in the series; 3 patients weremoribund on admission to the hospital, with typhoid



j.com/

Arch D


ecember 1977. D

ownloaded from

http://adc.bmj.com/

Diagnosis of acute bacterial pneumonia in Nigerian children 927

pneumonia, algid malaria, and congestive cardiacfailure respectively. One child died 6 days afteradmission after inhaling a peanut. No death was

related to needle aspiration; none had clinical evi-dence of pneumothorax or intrathoracic haemor-rhage. Necropsies were not carried out for religiousreasons.

Culture of pharyngeal secretions (Table 3). Potentialpathogens were isolated from pharyngeal secretionsin 45 (63%) of the 72 patients from whom throatswabs were collected. Normal nasopharyngeal com-

mensals were found in the remaining patients. How-ever, the degree of con,cordance between pharyngealand lung aspirate cultures was poor. Of the 53

Table 3 Comparison between pharyngeal culture andlung aspiration

Lung aspirate Pharyngeal culture

No. ofpatients from whom No. of patientspotential pathogens isolated from whom only

commensals isolatedIdentical with Different fromlung aspirate lung aspirate

Positive (n=53) 17* 18 18Negative (n= 19)- 10 9

*Of these, 15 were pneumococci.

patients in whom lung aspirate provided an aetio-logical diagnosis by microscopy or culture, an identi-cal organism, usually pneumococcus, was found inthe pharyngeal culture in only 17 (32 %). Cultures ofthe upper respiratory tract which yielded potentiallypathogenic organisms different from those isolatedin the lung aspirate (false-positive pharyngeal cul-tures) were obtained in 18 patients (34%). Falselynegative pharyngeal cultures (yielding only com-

mensal organisms) were obtained in 18 patients(34 %). 19 patients had negative lung aspirates andin this group potentially pathogenic organisms were

isolated from the upper respiratory tract in 10. Therelevance of these isolations is, of course, unknown.

CIE (Table 4). Sera were available from 45 patientsfor comparison with the lung aspirate cultures. In 7of the 29 patients with pneumococcal pneumonia(24%), pneumococcal antigen was detected in un-

concentrated serum. Pneumococcal antigen was

never detected in patients who had other organismsisolated from their lung aspirates. No proven case

of H. influenzae pneumonia yielded a positive resulton CIE, and there were similarly no false-positiveresults obtained from serum samples of patientsfrom whom other organisms had been isolatedfrom the lung aspirate.

Table 4 Comparison between countercurrentimmunoelectrophoresis (CIE) and lung aspiration

Organisms identifiedin lung aspirate

'CIE

CIE Pneumococcal H. influenzaenegative antigen antigen(no. ofpatients) identified identified

(no. ofpatients) (no. ofpatients)

Pneumococci 22 7 0H. influenzae 6 0 0Other bacteria only 4 0 0Negative 8 (+ 1)* 1 1 (+ 1)*

Note: Sera from 45 patients examined by CIE (4 patients had dualinfections).*Result of CIE on pleural fluid.

However, CIE provided the diagnosis in 3 of 11otherwise undiagnosed cases. The likelihood of thesebeing false-positive results is small since there wereno false-positive results in the patients in whom abacterial aetiology had been clearly established. In2 patients with pleural effusions, pleural fluid wasexamined by CIE and in one of these H. influenzaeantigen was detected in the absence of positivebacterial cultures.

Blood culture (Table 5). Blood cultures were avail-able in 36 patients for comparison with lung aspirates.

Table 5 Comparison between blood culture and lungaspiration

Bacteria Isolated No. ofpatients in whom blood culturefrom lung aspirate

Positive Negative

Staph. aureus 0 1Streptococci 0 3Pneumococci 4 17H. influenzae 0 5Coliforms 0 5

No organisms isolated 0 9

Note: Blood cultures were performed in 36 cases.

Only 4 of the 36 had positive blood cultures, allyielding pneumococci, and all having positive cul-tures for pneumococci on lung aspirate. These 4 hadthe classical radiological appearances of lobar pneu-monia. Thus, even in pneumococcal pneumoniablood culture was positive in only 4 out of 21 patients(19 %). There were no false-positive isolations fromblood culture, but the overall yield (11 %) was verylow.

Radiological changes (Table 6). Of the 43 patientswith lobar pneumonia, most (60%) had a puregrowth of pneumococci on lung aspirate. Of equalimportance, however, was the isolation of staphy-lococci, streptococci, and Gram-negative organisms



j.com/

Arch D


ecember 1977. D

ownloaded from

http://adc.bmj.com/


Table 6 Radiographic appearance of lungs in relationto aetiology

Organisms Radiographic appearanceIdentified inlung aspirate Bronchopneumonia Lobar Pleural

(n 44) pneumonia effusion(n=43) (n=13)

Pneumococci only 7 26 4Streptococci

only 4 1 2Staph. aureus

only 4 2 0Gram-negativerods* only 8 2 1

Mixedinfections 14 4t 3

No organismsisolated 7 8 3

*Includes E. coli, Klebsiella aerogenes, H. infliuenzae, Bacteroides spp.tThree patients had combined infection with pneumococcus andH. influenzae.

from 9 patients (21 %) with a purely lobar pattern ofconsolidation on chest x-ray.A radiological pattern of bronchopneumonia was

found in the remaining 44 patients. A wide varietyof organisms was isolated from this grout, Gram-negative or mixed infections being found in 50% ofthe patients. In contrast to the patients with lobarpneumonia, a pure growth of pneumococci wasisolated in only 7 patients (16 %). One patient withwidespread consolidation on chest x-ray was sub-sequently diagnosed as having bronchiectasis. Lungaspirate on this child yielded ,-haemolytic strepto-coccus group F. Radiological evidence of a pleuraleffusion was detected in 13 patients, yielding 10positive cultures, including a variety of organisms,but no staphylococci.

Overall rate of diagnosis (Table 7). The contributionsof blood culture and CIE to diagnosis were relativelysmall, direct needle aspirate of the lung providing adiagnosis in 79% of cases. In 21 patients lungaspirate (examined by Gram-stain and culture), CIE,and blood culture were all performed and a bacterialdiagnosis was obtained in 19 (91 %).

Table 7 Contribution of different techniques to diagnosis

No. of Pathogenpatients identified

Lung aspirate: Gram film 71 55 (77%)Culture 88 54 (61%)

CIE (serum) 45 9 (20%)Blood culture 36 4 (11%)

Any technique 88 73 (83%)

All four techniques 21 19 (91%)

Discussion

Needle aspiration of the lung. The technique of needleaspiration is not new (Leyden, 1883). In the earlydays it was severely criticized (British Journal ofChildren's Diseases, 1905), but has gained moreadvocates recently as its safety in children has be-come established. The technique has changed littlesince it was first described in the English journalsalmost 70 years ago, when lung aspiration was usedto produce a pure culture of the infecting organismsso that a specific vaccine could be prepared (Horder,1909). Some of the early studies were carried out totry to understand the natural history of pneumoniaafter acute infections (Ellison, 1931) and to study therelationship between the clinical and radiologicalfeatures of pneumonia and its aetiology (Lyon,1922). The procedure was ocasionally justified by theobservation that 'in some cases it appeared to hastenresolution' of the pneumonia (Ellison, 1931). Withthe advent of antibiotics, needle aspiration againbecame important as a means of determining thespecific aetiology before instituting appropriatetherapy (Disney et al., 1956; Klein, 1969; Hughes etal., 1969; Mimica et al., 1971; Davidson et al.,1976).Our study reaffirmed the value of needle aspira-

tion of the lung, permitting an accurate bacterialdiagnosis to be made in 79% of a group of childrenseverely ill with pneumonia. The safety of the tech-nique was amply demonstrated as only minor com-plications in 4 patients (5 %) were attributable to theprocedure.

Bacterial isolation rates from lung aspirates havevaried from 15% to 80%, being higher in the pre-antibiotic era (Ellison, 1931; Sappington andFavorite, 1936; Abdel-Khalik et al., 1938) and inpatients who had not received antibiotics (Mimicaet al., 1971). Mimica et al. (1971) obtained an isola-tion rate of 57% in untreated patients, but only 23%in those who had received antibiotics for more than24 hours before needle aspiration. In our study nopatient had received antibiotics during the currentillness.No deaths have been reported in childhood as a

result of needle aspiration of the lung. In 6 series ofneedle aspirations in children (Alexander et al.,1941; Disney et al., 1956; Schuster et al., 1968;Hughes et al., 1969; Klein, 1969; Mimica et al.,1971), complications occurred in 24 out of over 800children (3%). Of these only 2 were serious, bothcases of pneumothorax requiring drainage. Dick etal. (1974) reviewed 9000 needle aspiration biopsiesperformed in adult patients and reported 6 deaths.Some of these deaths have occurred in patients withmajor complicating factors such as atherosclerosis



j.com/

Arch D


ecember 1977. D

ownloaded from

http://adc.bmj.com/


(Pearce and Patt, 1974), or after excessive sedation(Meyer et al., 1970). Pulmonary hypertension wasthought to be a contraindication too. In one study,pulmonary haemorrhage occurred in thrombocyto-penic patients when a Silverman-type cutting needlewas used. A bevelled needle is relatively safe (Bandtet al., 1972), as confirmed by our study in children.

Bacterial isolations in our study were probablyaetiologically important, as cultures of needle aspir-ates in the lungs of 13 normal children by Mimicaet al. (1971) were all sterile. Any growth from aneedle aspirate of the lungs appears to be significant.

Culture of pharyngeal secretions. While accepting theaccuracy and safety of needle aspiration, it might beclaimed that there are other equally accurate andmuch safer methods of reaching a diagnosis. Evenin normal individuals, however, potentially patho-genic bacteria are usually present in sputum,pharyngeal secretions, and tracheal aspirate, and insome subjects in the main bronchi too (Laurenzi etal., 1961). Thus positive bacterial cultures of poten-tial pathogens from sputum or pharyngeal secretionsmight be expected in many patients, whatever thecause of their illness. Under these circumstancesbacterial pneumonia is likely to be overdiagnosedand overtreated while conversely the true diagnosismay be overlooked.Our study has confirmed the frequent irrelevance

of sputum or pharyngeal cultures in children, andprovides a warning that cultures from the upperrespiratory tract will frequently lead to a false aetio-logical diagnosis. The false-positive isolates frompharyngeal secretions obtained might have led to theuse of dangerous and expensive antibiotics requiringintravenous administration.Only the isolation of pneumococci in pharyngeal

cultures gave some indication of the true cause of thepneumonia; in 15 of the patients from whom pneu-mococci were isolated from the pharynx an identicalorganism was responsible for the pneumonia. Pre-vious workers have claimed that in most patientswith lobar pneumonia, pneumococci could be identi-fied in both sputum and lung aspirate (Sappingtonand Favorite, 1936; Alexander et al., 1941; Davidsonet al., 1976). However, in normal British children upto 45% may be harbouring pneumococci in thepharynx, while 36% may carry Staphylococcusaureus (Brimblecombe et al., 1958). Carriage rates ofpneumococci increase during minor respiratory tractinfections, when their isolation from the upperrespiratory tract is of dubious clinical significance(Brimblecombe et al., 1958).

Other direct techniques for obtaining respira-tory tract secretions (bronchial brush biopsy andtracheal aspirate) are technically difficult to perform,

may require a general anaesthetic, and are unprovenin children. In any case the known presence ofpotential pathogens in the lower respiratory tree ofnormal subjects would lend no more confidence totracheal culture results than to those of pharyngealor sputum culture (Laurenzi et al., 1961).

CIE. This is a rapid and sensitive technique for thedetection of bacterial antigen in body fluids (Lancet,1976). Species- and type-specific capsular poly-saccharide antigens are released during a variety ofbacterial infections (Lampe et al., 1976) and pneumo-coccal polysaccharide can be detected in concentra-tions as low as 0 5 to 1 [±g/ml of serum by CIE(Coonrod and Rytel, 1973). The technique is rapid,giving a result within an hour, and its specificity hasbeen amply verified.The contribution of CIE to the diagnosis of bac-

terial infections in our patients was small. The 24%rate of detection of pneumococcal antigen in theserum of culture-positive cases of pneumococcalpneumonia was slightly lower than that of otherstudies (Tugwell and Greenwood, 1975; Spencer andSavage, 1976; Michaels and Poziviak, 1976). Thereal value of the technique lies in its ability to detectantigen in the absence of any viable organisms. 3patients were diagnosed by CIE when cultures werenegative. Bacterial antigen has been shown to persistfor several days after the successful treatment ofbacterial pneumonia with antibiotics (Spencer andSavage, 1976), whereas bacterial cultures are notori-ously susceptible to prior exposure of patients toantibiotics. It is in these antibiotic-treated patientsthat CIE may prove of the greatest value. The rateof detection of pneumococcal antigen in lobar pneu-monia may be increased to about 50% when urine,concentrated by dialysis, is examined by CIE (Coon-rod and Rytel, 1973; Tugwell and Greenwood, 1975)and to 80 to 100% when pleural fluid is examined(Tugwell and Greenwood, 1975; Lampe et al., 1976).The presence of pneumococcal antigen in the

serum of a patient with a respiratory tract infectionimplies pneumococcal pneumonia (Tugwell andGreenwood, 1975; Spencer and Savage, 1976). Therelevance of antigenaemia in other bacterial infec-tions of the respiratory tract has not been thoroughlyevaluated.

Blood culture. Normally considered a routine investi-gation in childhood pneumonia, bacterial isolationrates by this procedure in classical lobar pneumoniahave only been 17 to 29% (Lyon, 1922; Sappingtonand Favorite, 1936; Mufson et al., 1967), similar tothe 19% obtained for pneumococcal pneumonia inthe present study. In spite of the almost certain



j.com/

Arch D


ecember 1977. D

ownloaded from

http://adc.bmj.com/


absence of prior antibiotic therapy in our patients,no organisms other than pneumococci were isolated.As special culture techniques were not employed, thisrate of isolation provides a realistic view of the placeof standard blood culture techniques in the diagnosisof bacterial pneumonia in children.

Radiological findings. It has generally been con-sidered that 'lobar pneumonia is caused by thepneumococcus', whereas bronchial pneumonia 're-presents a group of pulmonary infections' (Lyon,1922). Our study has emphasized the fact that acutesegmental consolidation (lobar pneumonia), thoughusually caused by pneumococcus, may be caused byother organisms. Previous studies carried out in theUK suggesting that lobar pneumonia may occasion-ally be caused by organisms other than pneumococci(Bath et al., 1964; Kinnell, 1973) have based theirbacteriological diagnosis on sputum culture and theirconclusions are therefore open to question. It isknown that the radiological appearance ofconsolida-tion may change during the course of disease; seg-mental lesions, occasionally seen early in the courseof staphylococcal pneumonia (Williams and Phelan,1975), may later progress to a more characteristicpattern with pneumatoceles and empyema (Disneyet al., 1956). It is therefore unwise to assume, evenwith the aid of sputum culture, that pneumococcalinfection is the sole cause of lobar pneumonia.The wide variety of organisms present in the group

defined radiologically as bronchopneumonia mayreflect the multiple predisposing factors (measles,whooping cough, malnutrition) in these patients (per-sonal observations). There were few distinguishingcharacteristics, some of which were misleading. Forinstance, 2 patients with pneumatoceles each hadpneumonia caused by multiple organisms, suggestingthat these may have been inhalation pneumonias(Gonzalez-C and Calia, 1975). The patients withpleural effusions again covered the spectrum ofbacterial diagnoses, except that no case of staphylo-coccal empyema was identified. Prediction of thecause of a pulmonary infection from the radiographicappearance was highly inaccurate in this study.

Conclusions. By applying all of the four techniquesavailable in this study (lung aspirate examined byGram-stain and culture, CIE, blood culture), it waspossible to identify a bacterial cause of pneumoniain 19 out of 21 patients. Whether the remaining 2patients had pneumonia caused by organisms whichwere undetected by our techniques (viral, protozoal,or mycobacterial) or whether this figure representsthe ultimate limitations of our techniques, is notclear. It must be borne in mind that this study wascarried out in an area where bacterial pneumonia

might be expected to be particularly prevalent andwhere antibiotic therapy had not pre-viously beengiven. A similar rate of diagnosis would not be ex-pected in countries where viral respiratory tract in-fections were more prevalent, or where antibiotictherapy was freely given to patients before admissionto hospital.However, the present study does have relevance to

the management of acute lower respiratory tract in-fections in children in situations other than thesavanna belt of West Africa. The technique of lungaspiration has been shown to be of value in theinvestigation of pulmonary tuberculosis (Schusteret al., 1963), pulmonary infections complicatingleukaemia (Hughes et al., 1974), diffuse noninfectivelung disease (Gellis et al., 1953), and in children withpneumonia unresponsive to normal therapy or withimpaired immunity (personal observations). A simplelung aspirate may safely provide a sufficiently largesample for bacteriology, virology, mycology, andcytology (personal observations).By the use of CIE, a simple and rapid technique,

a diagnostic tool is available whose results areindependent of the presence of viable bacteria inspecimens. Application of the technique to urineand pleural fluid, and the use of antisera to organismsother than the pneumococcus and H. influenzae,may improve its diagnostic yield.By the application of the techniques described in

this paper, it should be possible to identify the causeof bacterial pneumonia in children with the degreeof accuracy presently available in the diagnosis ofviral infections of the respiratory tract.

We thank Dr. B. M. Greenwood for performing theimmunological measurements, and the nurses anddoctors of the Emergency Paediatric Unit for theirhelp in the management of patients.

References

Abdel-Khalik, A. K., Askar, A. M., and Ali, M. (1938). Thecausative organisms of bronchopneumonia in infants inEgypt. Archives of Disease in Childhood, 13, 333-342.

Alexander, H. E., Craig, H. R., Shirley, R. G., and Ellis, C.(1941). Validity of etiological diagnosis of pneumonia inchildren by rapid typing from nasopharyngeal mucus.Journal of Pediatrics, 18, 31-35.

Bandt, P. D., Blank, N., and Castellino, R. A. (1972).Needle diagnosis of pneumonitis. Journal of the AmericanMedical Association, 220, 1578-1580.

Barrett-Conior, E. (1971). The non-value of sputum culturein the diagnosis of pneumococcal pneumonia. AmericanReview of Respiratory Diseases, 103, 845-848.

Bath, J. C. J. L., Boissard, G. P. B., Calder, M. A., and Mof-fat, M. A. J. (1964). Pneumonia in hospital practice inEdinburgh 1960-1962. British Journal of Diseases of theChest, 58, 1-16.

Brimblecombe, F. S. W., Cruickshank, R., Masters, P. L.,Reid, D. D., and Stewart, G. T. (1958). Family studies ofrespiratory infections. British Medical Journal, 1, 119-128.



j.com/

Arch D


ecember 1977. D

ownloaded from

http://adc.bmj.com/


British Journal of Children's Diseases (1905). Exploratorypuncture of the chest-abuse of the needle-a warning, 2,466-473.

Butler, C. D., Shaw, N. G., Hoffman, S. J., Ditkowsky, S.,McVey, E., and Zeldes, M. (1941). Pneumonia in children.A review of seven hundred and thirty-one cases. Journalof the American Medical Association, 117, 1840-1843.

Coonrod, J. D., and Rytel, M. W. (1973). Detection of typespecific pneumococcal antigens by counterimmunoelec-trophoresis. Journal of Laboratory and Clinical Medicine,81, 770-786.

Davidson, M., Tempest, B., and Palmer, D. L. (1976).Bacteriologic diagnosis of acute pneumonia. Journal ofthe American Medical Association, 235, 158-163.

Dick, R., Heard, B. C., Hinson, K. F. W., Kerr, I. H., andPearson, M. C. (1974). Aspiration needle biopsy of thoraciclesions: an assessment of 227 biopsies. British Journal ofDiseases of the Chest, 68, 86-94.

Disney, M. E., Wolff, J., and Wood, B. S. B. (1956). Staphy-lococcal pneumonia in infants. Lancet, 1, 767-771.

Ellison, J. B. (1931). Pneumonia in measles. Archives ofDisease in Childhood, 6, 37-52.

Finland, M. (1969). Diagnostic lung puncture. Pediatrics,44, 471-473.

Gardner, P. S. (1971). Acute respiratory virus infections ofchildhood. Current Problems in Clinical Virology, p. 32.Ed. by J. E. Banatvala. Churchill Livingstone, Edinburghand London.

Gellis, S. S., Reinhold, J. L. D., and Green, S. (1953). Use ofaspiration lung puncture in diagnosis of idiopathic pul-monary hemosiderosis. American Journal of Diseases ofChildren, 85, 303-307.

Glezen, W. P., and Denny, W. F. (1973). Epidemiology ofacute lower respiratory disease in children. New EnglandJournal of Medicine, 288, 498-505.

Gonzalez-C., C. L., and Calia, F. M. (1975). Bacteriologicflora of aspiration-induced pulmonary infections. Archivesof Internal Medicine, 135, 711-714.

Horder, T. J. (1909). Lung puncture: a new application ofclinical pathology. Lancet, 2, 1345-1346.

Hughes, J. R., Sinha, D. P., Cooper, N. R., Shah, K. V., andBose, S. K. (1969). Lung tap in childhood. Pediatrics, 44,477-485.

Hughes, W. T., Feldman, S., and Cox, F. (1974). Infectiousdiseases in children with cancer. Pediatric Clinics of NorthAmerica, 21, 583-615.

Hyde, R. W., Hall, C. B., and Hall, W. J. (1973). Newpulmonary diagnostic procedures. American Journal ofDiseases of Children, 126, 292-295.

Kinnell, H. G. (1973). Bacteriological findings in pneumonia.Lancet, 2, 919-920.

Klein, J. 0. (1969). Diagnostic lung puncture in the pneu-monias of infants and children. Pediatrics, 44, 486-492.

Lampe, R. M., Chottipitayasunondh, T., and Sunakorn, P.(1976). Detection of bacterial antigen in pleural fluid bycounterimmunoelectrophoresis. Journal of Pediatrics, 88,557-560.

Lancet (1976). Antigen detection in chest infections, 2,890-891.

Laurenzi, G. A., Potter, R. T., and Kass, E. H. (1961).Bacteriologic flora of the lower respiratory tract. NewEngland Journal of Medicine, 265, 1273-1278.

Leyden, I. (1883). 1Yber infectiose pneumonie. DeutscheMedizinische Wochenschrift, 9, 52-54.

Loda, F. A., Clyde, W. A., Glezen, W. P., Senior, R. J.,Sheaffer, C. I., and Denny, F. W. (1968). Studies on therole of viruses, bacteria and M. pneumoniae as causes oflower respiratory tract infections in children. Journal ofPediatrics, 72, 161-176.

Lyon, A. B. (1922). Bacteriologic studies of one hundred and

sixty-five cases ofpneumonia and postpneumonic empyemain infants and children. American Journal of Diseases ofChildren, 23, 72-87.

Meyer, J. E., Ferrucci, J. T., and Janower, M. L. (1970).Fatal complications of percutaneous lung biopsy. Radiol-ogy, 96, 47-48.

Michaels, R. H., and Poziviak, C. S. (1976). Countercurrentimmuno-electrophoresis for the diagnosis of pneumoccocalpneumonia in children. Journal of Pediatrics, 88, 72-74.

Mimica, I., Donoso, E., Howard, J. E., and Ledermann,G. W. (1971). Lung puncture in the etiological diagnosis ofpneumonia. American Journal of Diseases of Children,122, 278-282.

Morley, D. (1973). Paediatric Priorities in the DevelopingWorld. Butterworth, London.

Mufson, M. A., Chang, V., Gill, V., Wood, S. C., Romansky,M. J., and Chanock, R. M. (1967). The role of viruses,mycoplasmas and bacteria in acute pneumonia in civilianadults. American Journal of Epidemiology, 86, 526-544.

Nelson, W. E., Vaughan, V. C., and Mackay, R. J. (Editors.)(1969). Textbook ofPediatrics, 9th ed., pp. 42-45. Saunders,Philadelphia.

Nichol, K. P., and Cherry, J. D. (1967). Bacterial-viral inter-relations in respiratory infections of children. New EnglandJournal of Medicine, 277, 667-672.

Pearce, J. G., and Patt, N. L. (1974). Fatal pulmonary hemor-rhage after percutaneous aspiration lung biopsy. AmericanReview of Respiratory Diseases, 110, 346-349.

Rosenow, E. C. (1911). A bacteriological and cellular studyof the lung exudate during life in lobar pneumonia.Journal of Infectious Diseases, 8, 500-503.

Sappington, S. W., and Favorite, G. 0. (1936). Lung punc-ture in lobar pneumonia. American Journal of the MedicalSciences, 191, 225-234.

Schuster, A., Duffau, G., Nicholls, E., and Pino, C. M.(1968). Lung aspirate puncture as a diagnostic aid inpulmonary tuberculosis in childhood. Pediatrics, 42,647-650.

Spencer, R. C., and Philp, J. R. (1973). Effect of previousantimicrobial therapy on bacteriological findings inpatients with primary pneumonia. Lancet, 2, 349-351.

Spencer, R. C., and Savage, M. A. (1976). Use of counter androcket immuno-electrophoresis in acute respiratory infec-tions due to Streptococcus pneumoniae. Journal of ClinicalPathology, 29, 187-190.

Tugwell, P., and Greenwood, B. M. (1974). Bacteriologicalfindings in pneumonia. Lancet, 1, 95.

Tugwell, P., and Greenwood, B. M. (1975). Pneumococcalantigen in lobar pneumonia. Journal of Clinical Pathology,28, 118-123.

Williams, H. E., and Phelan, P. D. (1975). RespiratoryIllness in Children. Blackwell, Oxford.

Correspondence to Dr. M. Silverman, Departmentof Paediatrics and Neonatal Medicine, Hammer-smith Hospital, Du Cane Road, London W12 OHS.

Addendum

In a recently published paper (Chaudhary et al.(1977), American Journal of Diseases of Children,131, 902-907) an incidence of pneumothorax of32% was reported after needle aspirations of thelung in 202 immunocompromised children withpneumonia. The incidence was 43% in the 121children with Pneumocystis carinii pneumonia. Nodeaths were reported as a result of the procedure.



j.com/

Arch D


ecember 1977. D

ownloaded from

http://adc.bmj.com/

Documents

Diagnosis of acute bacterial pneumoniain Nigerian children · (ii) bronchopneumonia, in which there was more widespread or ill-defined consolidation. The x-rays of one patient subsequently